Optimization and Forecast Engineer - Energy Systems (f/m/d)

FION Energy; Berlin; Hybrid; EUR 90.000 - 120.000; Vollzeit

Benefits dieser Stelle

VSOP (virtual shares)

Berlin hybrid (remote within Germany)

Participation in company success

Zusammenfassung

FION Energy is seeking an engineer with deep experience in time-series forecasting and optimization to join as an early hire. You will work on forecasting and optimization models that predict industrial electricity use and PV generation, and determine battery operation against tariffs and markets, within physical and grid constraints.

You will work directly with the CTO, one of the co-founders, in a startup environment that blends strategic thinking with hands-on development.

Qualifikationen

• 5+ years of professional software engineering experience.
• Strong experience in time-series and probabilistic forecasting.
• Strong understanding of mathematical optimization (LP/MILP).
• Strong Python skills with the scientific and machine learning tool stack.
• Experience deploying forecasting/optimization models in a commercial software setting.
• Ownership mindset and platform-building drive.

Aufgaben

• Forecasting models: develop and improve time-series forecasting for industrial electricity consumption and PV generation, with probabilistic outputs.
• Battery dispatch optimization: design and improve models for scheduling dispatch across peak shaving, self-consumption, market trading, and grid constraints.
• Validation and monitoring: build simulation/benchmarking workflows and monitor forecasting/optimization model quality.
• Production integration: write and deploy production-grade code and integrate models into the core platform and edge devices.

Kenntnisse

5+ years software engineering

Time-series forecasting

Optimization (LP/MILP)

Python scientific stack

Forecasting/optimization model rollout

Ownership mindset

Tools

Python

PyTorch

Pyomo

OR-Tools

Jobbeschreibung

European industry is losing competitiveness because electricity here is more expensive and more volatile than in the US or China. The reason: renewables fluctuate strongly, factories consume constantly.

We close that gap. FION plans and installs the right battery system for an industrial site and runs it with AI in real time against tariffs and markets. The software that dimensions, controls and monitors those systems is our own, and it runs end to end: ML and optimization in the cloud, data streaming and control, and the edge device at the customer site. The result for the customer: significantly lower energy costs and measurable CO2 savings. We earn trust by delivering systems and operating them transparently.

We already work with more than 20 factories across industries such as plastics, automotive and food processing, and we are backed by early-stage investors.

We're looking for an engineer with deep experience in time-series forecasting and mathematical optimization to join us as an early hire. You'll work on the forecasting and optimization core of our platform: the models that forecast industrial electricity consumption and PV generation, and the optimization that decides how each battery operates and trades against tariffs and markets within physical and grid constraints. This is a startup role: you should be comfortable switching between strategic thinking and hands‑on work.

You’ll work directly with the CTO, one of the co‑founders.

Tasks

Forecasting models

Develop and improve time‑series forecasting models for industrial electricity consumption, PV generation, and other energy‑relevant signals, and extend them with probabilistic outputs. Evaluate forecast quality on noisy, incomplete, and non‑stationary industrial data, including its actual operational and economic impact.

Battery dispatch optimization

Develop and improve the optimization models that schedule battery dispatch across peak shaving, self‑consumption, spot market trading and flexibility marketing, while adhering to physical and grid constraints. Design how forecasts, uncertainty measures, physical constraints, and system states are used in downstream optimization workflows.

Validation, simulation and monitoring

Build and improve simulation, replay, benchmarking, and validation workflows to test model behavior before and alongside deployment in live systems. Build and use tools and processes to monitor and assess the quality of operational forecasting and optimization models.

Production integration

Design, write, test, and deploy production‑grade code for mission‑critical forecasting and optimization products. Improve robustness through plausibility checks, fallback behavior, re‑forecasting, and handling of low‑confidence or missing data. Collaborate closely with software engineers to integrate models into our core platform and edge devices.

Requirements

You bring:

• 5+ years of professional software engineering experience
• Strong applied experience in time‑series and probabilistic forecasting: forecast calibration, uncertainty evaluation, backtesting, and model validation
• Strong understanding of mathematical optimization, especially LP/MILP, and the ability to model real‑world systems through objectives, constraints, and operational rules
• Strong Python skills with the scientific and machine learning tool stack (pandas, NumPy, SciPy, scikit‑learn, PyTorch)
• Experience developing, releasing, and tracking the performance of forecasting or optimization models in a commercial software setting
• An ownership mindset - you want to shape the platform, not just implement tickets

Even better if you have:

• Energy domain knowledge - you understand power, energy, phases, and how the grid works
• Understanding of European electricity markets, flexibility services, and grid operation
• BESS‑specific experience: familiarity with BESS architectures, components, and operation, and direct experience with energy management algorithms for battery systems
• Experience with open‑source and commercial MILP solvers (e.g. Pyomo, OR‑Tools)
• Experience with optimization techniques such as stochastic or robust optimization
• Experience building forecasting and machine learning products in the cloud (e.g. AWS, GCP, Azure)
• Experience using Large Language Models to accelerate software development and reliably add new capabilities to real products
• German language skills (our customers and partners are German)

Ownership, not tickets

This isn’t a role where the models are decided and you fill in the blanks. The forecasting and dispatch optimization stack is greenfield - nothing exists yet, and you’ll design and build it from the ground up. At larger energy companies, this scope is split across multiple teams. Here, you own it all.

Real impact, fast

Your code runs physical batteries that save real factories real money. You’ll see results on a Grafana dashboard the same week you deploy.

Founding trajectory

As FION grows, you grow with it - from hands‑on builder to platform lead to head of forecasting and optimization. What takes 5 years at a large company, you’ll have from day one.

Compensation

• VSOP (virtual shares)
• you participate in the company's success
• Berlin hybrid (remote from within Germany possible for the right person)